agent-self-evaluation

Sort ECC components into DAILY vs LIBRARY buckets using repo evidence instead of guessing.

Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing.

Build persistent multi-agent operating systems on Claude Code with kernel routing, specialist agents, and file-based memory.

Engineering operating model for teams shipping with AI-assisted code generation.

Regression testing patterns for AI-assisted development to catch systematic blind spots.

Clean Architecture patterns for Android and Kotlin Multiplatform projects with layered module structure and dependency rules.